A safety-critical software reliability analysis method based on statistical model checking and bayesian theory
Zhanguo MA,Minjun PENG,Hidekazu YOSHIKAWA,Huayu YUAN,Bo YANG
DOI: https://doi.org/10.1299/jsmeicone.2019.27.1538
2019-01-01
The Proceedings of the International Conference on Nuclear Engineering (ICONE)
Abstract:The software in the digital reactor protection system is the safety-critical software. During the design and development, the safety-critical software goes through the strict test, review, verification and validation. The software products have prominent characters such as i) dedicated software; ii) designed defense in depth; iii) designed with fault tolerance techniques; iv) designed with fail-safe mode; v) have redundant configuration; vi) must have low failure rates. In addition, the software for the safety-critical system must be verified and validated independently and the software life cycle development process must be reviewed rigorously. Furthermore, extensive tests are performed for the software to not only remove the faults but also increase the confidence and reliability level of the software. Therefore, the quality of the safety-critical software is much higher than the commercial software. As well as, the safety-critical software failure date is scarce in the nuclear power plants. Hence, the methods that can be applied for the commercial software reliability analysis are not adapted for the safety-critical software reliability analysis. Considering the characteristics of the safety-critical software, a reliability analysis method that can be adopted to the safety-critical software is proposed based on the statistical model checking method. The proposed method can be applied to the early phase and the later phase of the software development life cycle. When the method is applied at the early phase, the software design schemes and the reliability allocation strategies are justified; when it is applied at the later phase, the probability failure on demand and probability of spurious logic trigger of the safety-critical software are estimated.